346 BOTANICAL GAZETTE [May 
form which he assumed to be O. grandiflora Aiton. It is evident 
that the first condition of success in such work consists in the 
purity and the immutability of the species which are to produce 
the hybrid. If they are already in a mutable condition, it is to be 
expected that their hybrids, or at least some of them, may com- 
bine the different lines of mutability of their parents; and at all 
events, the mutability of such a hybrid would be no proof that this 
phenomenon may be produced by means of crossing. On the other 
hand, if the species to be crossed, or even only one of them, were 
not pure, the hybrid might inherit this impurity and show phe- 
nomena which might easily be mistaken for mutations. 
It so happens that O. biennis is in a condition of mutability — 
analogous to that of O. Lamarckiana, although not developed to 
the same high degree. From time to time it produces dwarts, 
which are distinguished from it by exactly the same two characters 
which differentiate the dwarfs of O. Lamarckiana from their mother 
species, namely, low stature and sensitiveness to the attacks of 
some species of soil bacteria.2 Moreover, Stomps has shown that 
O. biennis may, although very rarely, double the number of chromo- 
somes in its sexual cells, which in O. Lamarckiana produces the 
two mutants O. gigas and O. semigigas.3 As is now generally 
admitted, O. gigas results from the pairing of two mutated sexual 
cells, each of which had a double number of chromosomes. 0. 
semigigas, on the other hand, is produced by the pairing of a 
sexual cell mutated in the same way, with a normal gamete; there- 
fore it possesses only 21 chromosomes (14+7), while the number 
in O. gigas is 28. As yet, only semigigas mutants have been 
observed coming from OQ. biennis, and it is obvious that the double 
combination must be much rarer. As a proof of this special kind 
of mutability in O. biennis, however, the observations of STOMPS 
are wholly sufficient. 
In quoting these facts, Davis says that if it can be shown 
tested strains of this biennis are able to produce new forms of specific 
?Stomps, Tu. J., Mutation von Oenothera biennis L. Biol. Centralbl. ctl 
535- 1912; also Zevisrra, H. H., Oenothera nanella De Vries, eine krankhafte Pian? 
art. Biol. Centralbl. 31:129-138. ro11. Vergl. ferner: Gruppenwelse Artbildung 
1913: 296-304. 
3 Stops, Tu. J., of. cit. p. 533. 
